In this episode, Jamey is joined by Karen Catlin, author of the Better Allies book series, for an in-depth discussion on the intersection of artificial intelligence (AI) and diversity, equity, and inclusion (DEI) in the workplace. After spending decades studying AI, Karen talks about how AI is shaping workplace practices, highlighting both its potential and the challenges it presents, particularly around bias, competency perception, and representation in AI driven projects.

They also talk about the competence penalty faced by underrepresented groups when using AI at work, the visibility and opportunity gaps in AI pilot projects, and the systemic barriers that persist even as organizations adopt groundbreaking technologies. The group also discusses prudent strategies for integrating AI responsibly, emphasizing transparency, diverse representation, and the need to continuously evaluate for unintended biases in AI-enabled systems.

I'm Jamie Applegate, Senior Director of DEI at EquityAtWork, and this is your DEI minute. Your go to podcast for leaders looking to navigate the ever evolving landscape of diversity, equity and inclusion in the workplace. Whether you're just starting out with DEI or looking to sustain your long term successes, each episode will provide you with the actions you can take to move DEI forward at your organization, all in 15 minutes or less. Join us every other week as we break through the noise and help you do DEI right. Let's get to it.

Before we get started. This is Michelle Pfefferman and I'm really excited to let you know that my new book, Do DEI Write is now available. This is your guide to the Equity at Work Maturity Model, which shows leaders how to make DEI part of every day and drive great results. You can get your copy through the link in the show notes or wherever books are sold.

Welcome everybody. On today's yous DEI Minute, we are very excited to welcome Karen Catlin, author of the Better Allies book series. Karen is a highly acclaimed workplace influencer, speaker and author. She's published four books. Her first was Better Allies Everyday Actions to Create Inclusive, Engaging Workplaces, and then followed it up with Belonging in Health Care, the Better Allies Approach to Hiring, and the Better Allies Way. She also emails a roundup of five ally actions to over 40,000 newsletter subscribers every week. I am one of those subscribers. It's wonderful.

She has a great community spotlight every week, so I encourage you to check that out. Previously, Karen spent 25 years building software products and serving as a Vice President of Engineering at Adobe. And during that time she witnessed a sharp decline in the number of women working in tech. And so she became frustrated but galvanized. She knew it was time to switch gears and focus on creating better workplaces where everyone can do their best work and thrive. So we are absolutely throw the haver on the podcast. Karen, welcome to your DEI minute.

Oh, it's a pleasure to be here, Jamie. Thank you so much for having me here.

Awesome. Well, Karen, obviously you are best known for the Better Allies book series. I have the book on my shelf over here and we at Equity Work recommend it all the time to clients and partners. The book is a phenomenal guide for anyone who has a stake in making workplaces more inclusive and engaging. In other words, everyone. And as soon as this episode ends, go order it. Just you'll be better for it. Please go order it.

And for today's conversation, we're actually going to go down a slightly different path and talk about artificial Intelligence or AI. So, Karen, what got you interested in AI and how does it connect to your work?

Oh, my gosh. I'm going to share. Often the hardest thing for a woman to talk about is her age, but I'm going to do that right now because it sets context. I first took an AI course 43 years ago, so. So I've been paying attention to AI a long time. I have a computer science degree. It was part of the curriculum. I studied AI back when I was in college.

So it was a long time ago, granted. And I can't say I ever really anticipated the tooling, the technology that we have today at our fingertips, literally. But I've been paying attention to it for a long time, and certainly a lot more in the last couple of years as I started understanding the start the intersection between artificial intelligence, the bias that can happen around that whole ecosystem, and the impact it has on our workplaces. And my focus now, after working in tech for 25 years, my focus now is all about how can we build better, better workplaces where everyone can do their best work and thrive, create the kind of workplaces that you as an individual want to work in and that others want to and need to have. So I started pay. It was just this perfect intersection for me of this background a long, long time ago, this more recent just curiosity and then realizing that there was an impact having on our workplaces. And I want to explore that, understand it, talk about it. So thank you for inviting me.

Yeah, no, I love it. And so in a previous episode of this podcast, I talked about AI sort of more broadly, sort of just general, like what they are, what LLMs are, and sort of how they work, how people can use them. And then we went on a couple use cases. But one thing I touched on, especially around some of the more problematic aspects of these platforms, is just the black box problem of these systems. So what we talked about was just that there's not really transparency into the algorithms that dictate what outputs we get. And so we have no insight, as you mentioned, sort of into the biases of the people who create these tools. We think of these tools as kind of this all powerful sort of mystical force. But somebody made that, and we have to think about, you know, who made it.

And so we treat the outputs that we get as very authoritative, sometimes a bit sycophantic, which I know is a complaint that people have about some of the chatbots. But it really means that we're not investigating why we're getting the responses. We're Getting. And so you mentioned a few things in a previous conversation we've had and also in an article that you wrote for HRM Outlook, which we will link to in the show notes, things you mentioned are things like a competence penalty for using AI, bias in AI driven performance reviews and access to AI pilot projects, which I thought were fascinating. So I was hoping you could share a little more about those things.

Yeah, definitely. So there are people who are studying this in an academic way, researchers out there, which I really appreciate when they share their work, publish it, and I've been just eating it up. So let's talk about that competency penalty. Turns out that when people use AI to do their job, they can face a competency penalty, meaning that they are seen as less competent. And this is especially true if that person using AI is from a stereotyped group. Stereotyped because they are underrepresented in their field, such as a woman in tech, my background, or an older worker in a youth dominated field of any sort. Those people, like everyone can face this competency bias and penalty, but when we're stereotyped we might face it even more. So just very quickly, let me see if I can summarize the research I've come across.

One is that doctors who use AI are seen as less competent, less empathetic and basically less skilled at their, at their job. So that's a concern because certainly there's a lot of good AI being deployed and helpful AI in the medical field for research as well as patient visits and all sorts of things. So it's a problem if you think, oh my doctor's using AI, I shouldn't be seeing this doctor. No, see, let's still see your doctor, right?

Go see the doctor.

Go see the doctor. Pay attention to your doctor. Secondly, let's talk about just tech first for a second and coding. There was a study done where people, participants in a study were shown identical snippets of code and were told a man wrote this, a woman wrote this, a man wrote this using AI, a man wrote this without using AI, a woman used this, wrote this code without using I or with AI. So the just the four kind of options. And then they were asked. The participants were asked to evaluate the code. They found that the people using AI they reported overall they were 9%, I think with less competent.

But there was a huge difference in the men versus the women. The men were only rated as 6% less competent when they used AI. The women were rated 13% less competent. So women were rated over twice as much less competent as men for using AI. So that's a problem. As we think about this. Technology is such a game changer. It is being sought after across tech companies as well as other companies that just use technology.

And if we're thinking that women aren't as good at their job because they're using this, this new technology, like, that's a big problem. Right. So those are just some examples of this competency bias that I want everyone to know about. So let's, let's amplify that, that research and make sure people have this on their radar as they are evaluating people.

Yeah, I really love that. And I mean, there's so many things, and I think I want to kind of drill down on that. I mean, what do you, what do you do about that? Like where women are kind of in a lose or anyone in from a marginalized background is kind of in a lose lose situation. They fall behind their peers if they don't use AI because everybody's using it. And I mean, there's the phrase like vibe coding. Even people who know how to code can whip up an app that basically mimics an SaaS platform. We're seeing massive changes in these things. And then so there's the pressure to use AI because everybody's using it.

But then, as you're saying, they get dinged and dinged much more harshly than their male peers and seen as less competent. So maybe there's pressure to not use AI to really show like, I know what I'm doing, you know, the old work twice as hard to go half as far. And, you know, so what are some potential solutions there? Both, maybe at the individual personal level for both men and women, but then also at the systemic level. So how do we really combat that idea where we're saying, you know, we shouldn't be dinging anyone more harshly based on the strategies they're using.

I know, I know. Well, I want to come back to your question. And so, and the reason I want to come back to it is there's something else underlying that I have learned in that women are underrepresented in their knowledge of generative AI. Specifically, there was a study by Boston Consulting Group that studied this across their client base. And junior women were less knowledgeable about generative AI, systems, tools, technologies and how to apply it. And the research chalked it up to two things. One is the women lack the informal networks, maybe where this stuff is being discussed and talked about. So that's a problem.

And we know in a DEI space when we have most of us have just like me networks where they're filled with people just like us because we like hanging out with people like us. And if women aren't in the networks where the AI is being talked about in the rooms where it's all happening, you know, that's a problem. But the second thing this research pointed to is that women were underrepresented in AI pilots. So the pilots being, let's try this out in this functionality in this part of the organization with this client, whatever it might be, women are underrepresented. And again this, this is frustrating because this is a, this is not a new problem where women are underrepresented in pilots or stretch assignments. Stretch assignments that often lead to great career growth because they're high profile. You know what I'm talking about, High profile, high visibility. Executives, leaders care about this stuff and they're often looking at the results and getting to know the people working on it.

So when the women are underrepresented in any kind of pilots, but especially AI pilots in this, in this generation that we're in right now where it's AI everything that's going to impact their careers short and long term. So I just needed to like, also like weave that into the conversation here, is that there's a lot going on. It's the competency bias as well as a visibility bias that's, that's happening. So, so back to your question though. What can we do on a personal level or systemic level? Personal level. Like if you are by any chance working on an AI pilot or assigning those AI pilots, take a, just take a pause, take a beat and think about who are my go to people that I am thinking about tapping to join this pilot to be a part of it. And if it's sort of the same people you always tap or the people who kind of like remind you of your younger self or whatever that might be, or the group is not representative of your employee base and have some diversity in there, just take a pause, think about who else could be a valuable person, expand your network, expand the group that you are recruiting from, so to speak and look to create more diversity in your AI pilots. So that's going to be good for the, you know, that short term and of course we know.

I, I'm sure you know, Jamie, like just, it's a DEI well known fact that when you have diversity on any kind of team, you're going to lead to more innovative results. So this is also just good from. Let's make sure this AI pilot is going to be successful. Okay, so I'll mention the second thing I want to mention and I love this. After I shared the research about that coding competency penalty in my newsletter. You know, I hear from a lot of subscribers every time I send out a newsletter. I love it. But one person reached out and said, you know, because you told me about that competency penalty, I'm now going to flip it around at my organization and I'm going to change our job ladder.

The descriptions of the responsibilities and expectations of each level of the job ladder to have an AI competency competency related item in that so that it shows that AI is celebrated at every level and what that expectation is around not just literacy but competency in using AI. So I loved that and I think that's a best practice we can do if we are in charge of a system like that, that we can influence how the job ladders are described and implemented. Yeah, I don't know. Jamie, you spent a lot of time thinking about this too. Would you say there's anything else we can do to address this competency?

I mean, I'm sure there are so many things, but I think those are great places to start. Just that question of like who's not looking around the room and saying who's not here, who's not represented here? I don't, it's not a game because I don't play it as a game, but I do a thing where I read too much about artificial intelligence from every angle. Both trying to understand the technology underpinning it. I am not didn't come up through tech like you did, so I don't have that sort of background in it, but so trying to understand it from the technical side and then also just a lot about the potential ramifications and the consequences of it. And I do a thing where I notice who do they ask for expert opinion and I try to understand and then I look those people up and it's rare that the person that they are talking to as a stated expert is a woman. And that's not to say that they are not women experts. They're not even being highlighted here. And so it is this sort of becomes a systemic issue where if you look at the founders and the leaders of the most well known AI firms, all men.

And I think you're absolutely right that there is that who's like us kind of mindset. And so I think asking who's in the room and sort of making that sort of personal or company wide pledge and to say like we need this to be representative because the AI is only going to get stronger through that and be actually more useful to the people using it. And then I really do love the idea of saying, I mean, we talk about this with clients and partners and all the time of just. And it's a DEI issue, but it's also not just a workplace issue. Is what are the expectations laid out for every role? Do we have clear role expectations? And when it comes out to rolling out AI and saying we're going to adopt AI. And I live here in Pittsburgh, and Duolingo is sort of a crown jewel in Pittsburgh's burgeoning tech scene. It was founded, I think, from Carnegie Mellon, and everyone loves the little duo owl. He's cute.

And they sort of articulate. They're an AI first company. And they ask on job applications, how do you foresee using AI in this role? And what's. You know, and I imagine they have this internally. I mean, I don't work there, so I don't know. But it is. What is the expectation? What does it mean to be competent? What does it mean to utilize AI for this role? What level of skill do you need with AI to be able to advance? And really sort of outlining that for the role instead of saying these sort of very broad statements of we're going to be an AI first company. Well, what does that mean? And what does it mean to do it responsibly? So I.

No, I think those are a great place to start. And I think, I think also companies. I mean, I guess another one I would say is companies should be really willing to articulate what they know and what they're thinking about AI very transparently, because I think it scares a lot of people, especially from like a baseline, am I going to have a job to be able to get my needs met? But also from an economic system level, are we going to have mass layoffs?

I'll also add there's so many people who are rightly concerned about the environmental impacts of AI and these data centers that are massive and are energy consumers and affecting water level. Table. Water table levels and, you know, so people are concerned about that. People are concerned about, is it eroding critical thinking skills as. As we start relying on it more and more, is it eroding creativity and so forth. So there's a lot to be concerned about. And at the same time, it's an incredibly powerful tool when used responsibly.

Yeah, I think I saw something recently. It was like images of brain scans and I mean, from initial glance they looked very different, but it meant nothing to me until I read what they actually meant. But one was basically saying when using AI to answer something versus kind of noodling on a challenge, different parts of your brain lights up and they're starting to do this research that I find fascinating and so valuable for just understanding because we kind of initially think about it in the immediate of how's it sort of changing the workforce, but how's it changing education? How is it changing, you know, what children are experiencing? And, and then, I mean. And yeah, the environmental stuff is also huge. I mean, I will, I will say from a wider lens, as it really say, I will admit I fall on the neutral to cynical side of things. I think that it is an incredibly important thing that is growing and changing. And I don't think we fully, I think, you know, there's the people on the super doomsday side that this is going to, you know, end all life on earth in the next 10 years. And then there's the people sort of on the, you know, the, the evangelists who are like, we're going to have infinite money and infinite everything and everybody's going to give it.

I was like, I mean, I was like, I thought, okay. And so I sort of am like, I think it's going to be somewhere in the middle where I think it's going to be a critically valuable thing, but I think there need to be a lot of guardrails put around it. I mean. Yeah, so I'm certainly not an evangelist.

Yeah, yeah. And, and I think that middle ground, we'll call it, if that's okay to say where you are is where I think a lot of people should be who are in workplaces today. Middle ground, being supportive but skeptical at the same time. And I think from a better allies point of view, understanding that there are all of these concerns, whether you personally have them. I don't live near a data center. My electricity is fine, whatever. You might not personally have these issues, but hearing those from colleagues and realizing this is a concern, or any other issue that might be a concern about AI from a better allies point of view, I do want to encourage people to be asking the tough questions of do we even need to use AI here? Is it helpful? Is it worth the, the not the risk even, but is it worth the impact that we will see down the road or as a result of this? So I think we need to have this kind of skeptical supporters.

There was a really, I'll say as I, after I read the HRM Outlook article you wrote. I was sort of at the bottom of it and there obviously links to other relevant articles. And one was, I can't remember who wrote it. I hope that the link is just evergreen there. But it was saying, you know, I didn't, you know, didn't free up 20% of your cost. It free up 20% of, like, productivity for workers. And it was a very sort of saying, like, you should implement AI as a tool to support productivity, not as a genuine replacement. It can replace certain tasks that maybe take a long time.

There should always be that human oversight. And it's sort of change that. And I do think an important thing is to live in the middle and say, just to be thoughtful and considerate about it, of saying, you know, how do we. Are we actually thinking through all of the consequences of this? Because if I think if the evangelists and the people running these companies tend to be like, they're very bullish and they're like, we're just going to run to the future and it's going to be beautiful. And I was like, I don't think people. I don't. I think there's going to be a lot of destruction on the way toward that beautiful city on a hill that they're talking about. But I think, you know, I guess as we sort of get to the end of our conversation, I mean, how, you know, I do, yeah, I do sit on that sort of neutral, the cynical kind of cautious side.

And I mean, I work in the consulting field, which is always. One of the ones that is mentioned is like, this is one of the first to go. So I'm always curious to hear, you know, how. How do we do it responsibly? What are some of those, you know, is it necessarily a massively disruptive force? Can it be something else? Something less destructive, more constructive? Are there ways to mitigate some of those really scary things about AI? I'm thinking about companies that are already announcing, I know block announced a 40% layoff to say that they're going to be AI first. And that kind of sent, you know, shockwaves through the tech community and the larger community of like, they're really doing it, you know, what. How do we do it responsibly?

Yeah, yeah, yeah. And, you know, I, I don't want to comment on Block's business model. I don't know, you know, I don't

know enough about it. I just know that, yeah, we announced the layoffs and they said it was because they're going AI And I was like, yeah. And again, without a ton of extra information, we're sort of left. If you were sort of an AI doomsayer, you were like, they're doing it. And then if you're an AI evangelist, you're like, look how amazing they're, they're already able to do this.

Right, Right. And there's always another narrative too, of always. It's a convenient scapegoat to, to, to name at this point. In fact, it's so convenient, it's also really powerful with investors and Wall street and everything else. So we don't, we, we don't know specifically about that. And I don't want to comment yet on the impact on the workforce, on workforce sizes and layoffs and stuff like that. I think it's, it's all kind of emerging and people are figuring this out and trying to understand it all. But what I do want to talk about, you know, and you asked you how can we do use AI responsibly? Sorry.

I do want to comment though on your question about how do we do AI responsibly. I think that there are a lot of mandates now to use AI. And every time you're using AI in a new setting, think about what could possibly go wrong. Basically, you now have an AI powered performance management system to do quarterly reviews, performance talent management kind of thing. What could possibly go wrong? Well, what could possibly go wrong is that the reviews that are coming out of that system that are taking all the data that was collected could be creating a biased, more critical output for certain members of, of the, the team that you're working with. There's pro, there's proof about that there could be bias coming in because people are now using bias to come. Or, excuse me, there could be bias coming in because people could be using AI to create their 360 comments for their manager or their teammates. And there's an interesting experiment by Dr.

Kieran Snyder who found that you can prompt an AI chat GPT with the same question and just change the pronouns and the feedback that comes back for a woman versus a man, the woman's is longer and more critical in nature. It's fascinating. So assume that there could be bias at every point. The stuff coming in to this performance management stuff system and the stuff coming out. And that's just one example. So wherever you're using AI, think about what could be the potential impact, what could go wrong here, especially from an inclusion lens. And let's make sure that we are asking the tough questions, questioning what could be happening and that's the first step and that's something we each individually can do.

I love that. I mean, I, I definitely, I'm a very like, operational thinker. I like to think through like, what is the system and how are we moving through this process? I love processes and plans and everything in its right place. And I love the idea of that. And I think that it's, yeah, it's think through that and then I think it's, you know, try it out and then also do a little bit of a post, you know, post review of saying, how did it go? Did we see, you know, have, have those questions already in mind of like, are we looking, are these comments equally valuable to all groups? Are we noticing trends of like, oh, women got dinged a lot on these performance reviews? Or oh, this group got dinged a lot or this group got, you know, a little bit lower? If you're doing like a five star, five, five point scale, why did this group actually, is this group actually like a 2.7 on average or what's going on? They're being willing to kind of take the time to excavate and slow down. And I think. But the biggest thing, yeah, I think is it seems as though, you know, the, the urge is definitely to sort of adopt quickly and move. And I, from what you're saying, I think maybe the more prudent thing is to just slow down a little bit.

Take a moment before, during and after using AI to say, like, this can be a really valuable tool, but what's the best way to do it in a way that's really going to advance our workplace and not like, cause unintended damage. So I really, I really love that.

Yeah, 100%, I love it.

Well, I think that is a great place to wrap up. And so I will say after this conversation, I feel significantly less Doomsday esque about AI and I think there are ways to do it responsibly. So thank you, Karen. And so you all can pick up Better Allies wherever books are sold. You can sign up for the Better Allies newsletter@better allies.com and you can connect with Karen on LinkedIn, follow her on Instagram and YouTube @better allies. She also has an article out with more to come about AI. She's really getting into that space and how it's being implemented in ways that can harm or advance inclusion, equity and engagement. So we'll link to some of those in the show notes and we encourage you to check those out.

Karen, thank you so much for joining us and for your leadership in this work.

Thank you. It was a pleasure speaking with you today. Take care.

That's a wrap. I'm Jamie Applegate, and that's your DEI minute for today. Thank you for listening. Please be sure to follow us wherever you listen to podcasts. And don't forget to leave us a review. If you ever have questions, please visit our website or send us an email. You can also sign up for our newsletter and follow us on LinkedIn, YouTube, Twitter, and Instagram. Links to everything can be found in the episode Notes.

This episode was edited and produced by Podgrove with podcast art by me, Jamie Alpine,

Sam.